Objectives To evaluate the association of the use of individual immunosuppressive agents and mortality in a longitudinal cohort of SLE patients

Methods Patients who fulfilled ≥4 1997 ACR criteria for SLE and followed in our rheumatology clinics between 1995 and 2014 were longitudinally followed. Data on demographic characteristics, clinical manifestations and ever use of immunosuppressive agents were retrieved from our database. A propensity score for the use of individual immunosuppressive agent was calculated for each patient based on the odds ratios obtained from separate logistic regression models, with different drugs being the outcome variables. Cox regression models were established to study the effect of the use of various immunosuppressive agents and mortality of our patients over time, with adjustment by the propensity score, age at onset of SLE and sex. For those who died during the disease course, data were censored at the time of death and for those lost follow-up, data were censored at the time of their last clinic visits.

Results 803 SLE patients were studied (742 women, 92%). All were ethnic Chinese. The mean age of onset of SLE was 33.2±14 years and the mean follow-up time of the entire cohort of patients was 10.8±7.7 years. The cumulative manifestations of SLE, in decreasing order of frequency, were musculoskeletal (70.3%), renal (52%), facial rash (46%), leukopenia (36%), photosensitivity (26%), thrombocytopenia (25%), hemolytic anemia (21%) and serositis (19%). The frequencies of patients who were ever treated with various immunosuppressive drugs were as follows: prednisolone (85%), azathioprine (AZA) (63%), cyclophosphamide (CYC) (25%), mycophenolate mofetil (MMF) (27%), calcineurin inhibitors (CNI) (23%) and hydroxychloroquine (HCQ) (69%). 97 patients (12%) died (mean time to death 9.8±8.0 years) and 56 (7%) patients were lost to follow-up. The commonest causes of death were: infection (44%), cerebrovascular events (12%), cardiovascular events (10%), malignancy (8.2%), pulmonary hypertension (7.2%), interstitial lung disease (3.1%) and uncontrolled SLE activity (3.1%). Cox regression analyses adjusted for age, sex and vascular risk factors (current/past smoking, hypertension, diabetes mellitus, LDLc/HDLc >3.0) revealed that the ever use of MMF (hazard ratio [HR] 1.89 [1.19-3.02]; p=0.007), CYC (HR 1.67 [1.04-2.68]; p=0.04) and HCQ (HR 0.52 [0.35-0.79]; p=0.002) was significantly associated with mortality. Conversely, the use of CNI (HR 1.10 [0.67-1.82]; p=0.71), AZA (HR 0.98 [0.64-1.50]; p=0.92), or prednisolone (HR 1.06 [0.51-2.01]; p=0.85) was not significantly associated with SLE mortality. After further adjustment for the propensity scores, it was demonstrated that the ever use of HCQ (HR 0.53 [0.34-0.84]; p=0.006) and AZA (HR 0.55 [0.34-0.89]; p=0.02) was significantly associated with a reduction in SLE mortality (47% and 45%, respectively). The use of MMF (HR 1.42 [0.87-2.32]; p=0.17) and CYC (HR 1.24 [0.75-2.04]; p=0.40) was no longer significantly associated with an increase in mortality.

Conclusions In this longitudinal cohort of Chinese SLE patients, the ever use of HCQ and AZA was significantly associated with lower mortality. However, treatment with prednisolone, CYC, MMF or the CNIs was not associated with a benefit of survival.

Disclosure of Interest None declared